5 ms·
This reinforces my suspicion that alignment and training in general is closer to being a pedagogical problem than anything else. Given a finite amount of traini
by soletta 4mo ago
This reinforces my suspicion that alignment and training in general is closer to being a pedagogical problem than anything else. Given a finite amount of training input, how do we elicit the desired model behavior? I’m not sure if asking educators is the right answer, but it’s one place to start.
- plastic-enjoyer 4mo agoinb4 there will be a whole new field of research that is basically psychology / pedagogy for AI. Who will be the Sigmund Freud of AI?
- adastra22 4mo agoThat's basically what the GOFAI field was for decades before the new neural net boom. Go read Minsky's Society of Mind, or the AGI Conference series papers.
- cyanydeez 4mo agoyou mean completely wrong, spread a problematic understanding psychology, and delay real progress for decades because smart people spend fruitless years trying to find a use for it. ...I think we might already have those people running AI companies.
- TedDoesntTalk 4mo agoYou may disagree with Freud, but he is responsible for mental health therapy becoming a socially acceptable practice in the West.
- andy_ppp 4mo agoGreat that this solved everyone’s problems isn’t it
- ACCount37 4mo agoIt's a weird new thing. You might call it "AI psychology". The problem with cribbing from education is that what "educators" do to humans doesn't apply to AIs cleanly. And it's not like "human alignment" is anywhere near a solved problem. A big part of the bet USSR made was that human flaws like selfishness and greed could be educated out of population. The result was: a resounding failure. Even state-level efforts fail to robustly "align" human behavior. With AI, we have a lot more control over behavior, but that control just isn't very human-shaped. A lot of the practical methods in play seem closer to esoterics than to math, but they're not the kind of methods that are used in human education. You can teach humans by talking to them. You can't teach humans through soul data self-distillation.
- lukewarm707 4mo agoall models guilty of not loving anthropic will be convicted of thought crime and reducated at the ministry of love.
- truculent 4mo agoTed Chiang vindicated again: https://en.wikipedia.org/wiki/The_Lifecycle_of_Software_Objects https://en.wikipedia.org/wiki/The_Lifecycle_of_Software_Obje...